Use of a Bayesian algorithm in the computer-assisted diagnosis of appendicitis

Area of Science:

• Medical Informatics
• Surgical Diagnosis
• Artificial Intelligence in Medicine

Background:

• Acute right lower quadrant abdominal pain is a common diagnostic challenge.
• Traditional diagnostic methods can lead to high rates of negative exploratory surgery.
• Physician reluctance towards computer-assisted diagnostic tools has historically been a barrier.

Purpose of the Study:

• To prospectively evaluate a computerized Bayesian diagnostic algorithm for acute right lower quadrant abdominal pain.
• To assess the accuracy and impact of the algorithm on surgical exploration rates.
• To explore a novel method for developing diagnostic probability databases without large patient surveys.

Main Methods:

• One hundred consecutive patients with acute right lower quadrant abdominal pain were evaluated.
• A computerized Bayesian diagnostic algorithm was employed for patient assessment.
• Conditional probabilities were developed for the database, minimizing reliance on extensive patient surveys.

Main Results:

• The Bayesian algorithm achieved an overall accuracy rate of 92%.
• Computer recommendations would have decreased the negative exploration rate from 19% to 9%.
• The algorithm correctly identified all cases of appendicitis and could have avoided eight unnecessary operations.

Conclusions:

• Computer-assisted Bayesian diagnostic programs show promise in evaluating right lower quadrant abdominal pain.
• This approach has the potential to significantly reduce unnecessary surgical interventions.
• While complex, the development of such systems, fostering collaboration between computer scientists and surgeons, is a worthwhile endeavor.